With the ascendancy of "evidence-based medicine," much attention has been paid to the quality of the evidence, with randomized controlled tdals (RCTs) considered the "gold standard" against which all other types of studies are compared. There are, however, numerous clinical areas where the initial evidence is provided by observational studies such as prospective cohort and case-control studies, and only later confirmed or refuted by a randomized trial. We plan to examine these topics, and look specifically at predictors of the "best" answer among the observational studies. Specifically, we hypothesize that a priod identification of biases in an observational study will be related to systematic differences in its results. As a further step, we intend to establish the directionality and potential magnitude of the biases by comparing the results of observational studies to the results of "gold standard" RCTs performed on the same clinical topic. We will use meta-analyses of observational studies to identify topics where a large number of observational studies exist (often with some degree of heterogeneity), and compile information on the potential types of bias present in each study. To date, over 20 such areas with numerous observational studies and a gold standard RCT on the same clinical topic have been identified. Our research could identify potential deficiencies in the available data for a given area and guide future investigators on what common pitfalls to avoid when conducting original research in that field.